Simplicity.

Keeping the intervention simple is widely considered to be a predictor of success [12]–[14]. The key to “rapid and massive scale-up” of antiretroviral therapy (ART) for people living with HIV in Malawi, say Harries and colleagues, was to “keep the principles and practices of ART delivery as simple as possible” [12]. Billings and colleagues argue that post-abortion care was successfully scaled up in Bolivia and Mexico because it met the criteria for scalability—cost-effectiveness, simplicity, and replicability [13]. A World Bank review of scaling up rural development interventions found that strong efforts at simplifying models or programs were associated with success [14].

One interviewee, who previously led a multilateral health organization, said: “If the intervention is simple, agreed, and there are no dissenting views, scale-up is much more likely to happen.”

Scientifically robust technical policies.

Technical experts who have managed large-scale implementation argue that getting the technical policies scientifically robust before going to scale was crucial for success [15]. For example, before directly observed therapy short course (DOTs), a treatment program for tuberculosis, was scaled up across India, “all technical policies and detailed training modules for every level of staff were written, extensively revised, field tested over a period of several years, finalized, and disseminated widely” [15].

Strong leadership and governance.

Case studies in the literature and interviewees' experiences of successful scale-up suggest that strong leadership played an important role [5],[16]. For example, a leadership development program in rural Egypt was associated with an increase in the number of new family planning, prenatal, and postnatal visits and a fall in the maternal mortality rate [16].

Several interviewees pointed to the important role of leadership in Uganda's success in scaling up HIV interventions [17]. The head of an African medical school argued: “Leadership is critical—at national level, regional level. Uganda's success in scale-up was because there was committed leadership to scaling up.”

Engaging local implementers and other stakeholders.

A recurring theme among those I interviewed was the importance of getting buy-in from local implementers and other key stakeholders. The former director of a global health initiative (GHI) highlighted the role of local medical associations: “In neonatal health, if we don't have pediatric associations on board, forget it! For oral rehydration solution, after it was unequivocally demonstrated that it saved lives, the WHO and UNICEF were still doing local trials, randomized controlled trials. These didn't have scientific value but it led to buy-in locally.”

Using both state and non-state actors as implementers.

Non-governmental organizations (NGOs) have played a crucial role in successful scale-up in many settings [5],[6]. A professor of global health said that an NGO in Bangladesh, the Bangladesh Rural Advancement Committee, is now reaching more people with health interventions than the government is reaching. A former director of a GHI discussed how non-governmental recipients of support from the Global Fund to Fight AIDS, Tuberculosis and Malaria have performed better than governmental recipients: “They [NGOs] are more nimble, there may be less corruption, they're quicker off their feet, individuals [in NGOs] are more motivated. From 2005, there are compelling data to show non-governmental recipients did better.”

Interviewees also discussed the role that private providers can play in scaling up, complementing the role of state actors. An expert in health systems reform discussed the importance of removing barriers in both sectors: “For a market solution: are you allowing entry to the market? For the public sector: are you getting tied up in bureaucracy?”

Applying diffusion and social network theories.

Rogers identified five factors that are positively associated with the faster diffusion of an innovation (Table 1), five types of adopters (innovators, early adopters, early majority, late majority, and laggards), and five stages in the adoption of a new innovation (awareness, interest, evaluation, trial, and adoption) [18]. Applying such diffusion theories at the right time, in the right place, within a health system has been a factor in the success of several LMIC scale-up programs [5],[19],[20]. McCannon and colleagues argue that successful scale-up programs have also paid attention to “the nature of the social network into which they wish to disseminate new practices,” for example, by considering how people interact and who are the early adopters [5].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Factors associated with faster diffusion of an innovation.

https://doi.org/10.1371/journal.pmed.1001049.t001

Cascade and phased approaches to scale-up.

A national scale-up program for adolescent reproductive health services in health facilities and schools in Tanzania used a “cascade model”—regional trainers supervised district trainers, who in turn trained teachers and health workers [21]. The model succeeded in scaling up the intervention to 75%–100% of intended schools and health facilities in four districts of the country.

A related concept is the notion of going to scale in a phased manner, beginning with a pilot program, followed by step-wise expansion, learning lessons along the way to help refine further expansion [22]. Case studies of successful “phased” scale-up include Thailand's “100% condom” program [23], Guinea's national program to scale up HIV prevention and treatment tools [24], and scale-up of a harm reduction program in Asia [25].

Tailoring scale-up to the local situation, and decentralizing delivery.

A variety of successful scale-up projects tailored implementation to local conditions on the ground and decentralized delivery so that clinics were closer to the target communities [11],[21],[26]. One interviewee, the expert in health systems reform, said: “You must tailor the message to local circumstances. You need some capacity—for example, a toolkit, a support network—to help local decision-makers to interpret a national decree.”

Adopting an integrated approach to scale-up.

While many of the most high profile scale-up campaigns have been vertical in nature (e.g., scaling up insecticide-treated bed nets or ART), a complementary theme in the literature on successful implementation is the value of integrating scale-up activities into existing health systems [24],[27]. Using existing systems can be valuable if rapid scaling up is important, such as using established food and beverage delivery systems to rapidly distribute condoms [27]. HIV and mental health services have been scaled up in LMICs by integrating them into existing health services [26],[28]; similar integration is being planned for cancer services [29].

A related approach is to deliver disease control tools in an integrated manner, such as integrating child survival strategies with immunization [30], linking HIV and tuberculosis services [31], and integrating the control of different neglected tropical diseases [32].

An engaged, “activated” community.

Scaling up by engaging community members or community health workers is well described in the literature. Examples include the Bangladesh Rural Advancement Committee's engagement of itinerant health workers [5] and Pakistan's “Lady Health Workers” program [33]. The active participation of the community in planning, implementing, and monitoring interventions is widely cited as a crucial factor in successful scale-up [11],[21],[24],[34].

A recent multicenter study found that a community-directed intervention strategy was more successful than other delivery strategies for scaling up distribution of ivermectin (an antiparasitic medication), vitamin A supplements, and insecticide-treated bed nets [34]. Recent evaluations of community health worker programs have shown a positive impact on child health [35], while a recent randomized trial in Bangladesh found that scaling up community-based participatory women's groups was associated with reduced neonatal mortality [36]. One interviewee, a former director of a GHI, pointed out that community engagement also means engaging patients: “If patients are not involved, you can't implement. In AIDS, it's the rule. Patients are experienced experts. They've gone through it. They know.”

Political will and national policies.

National, regional, and local policy commitment are cited as important factors in the successful scale-up of breastfeeding in many LMICs (e.g., Bolivia and Madagascar) [37], post-abortion care in Guatemala [38], and DOTs in India [15]. Several interviewees said that political will has played a crucial role in scaling up adult male circumcision in some countries (e.g., Kenya), condom distribution in Thailand, and needle exchange in Iran. “Uganda has not been successful in scale-up of male circumcision,” said the head of an African medical school, “because there's no national policy yet. WHO recommends it should be scaled up. Therefore a policy or guideline must be in place—that's a success factor. Countries with a policy are making greater strides.”

Country ownership.

Several interviewees discussed the importance of country ownership—and of moving away from traditional donor-recipient relationships in which donors dictate the terms—in the success of national scale-up programs in Africa. A former GHI director said that countries that had demonstrated successful scale-up through support from the Global Fund to Fight AIDS, Tuberculosis and Malaria were “special countries where the antipathy for neo-colonial relationships, the national pride, the desire for a new relationship between poor and rich countries overwhelmed the comfort with old, cozy donor relationships.”

Incorporating research into implementation (“learning and doing”).

Simmons and Shiffman argue that successful scale-up “requires the systematic use of evidence to guide the process and incorporate new learning” [11]. India's successful program to scale up DOTs adopted such an approach—monitoring was accompanied by timely feedback to implementers. Incorporating research into scale-up is also important for testing the transferability of a successful pilot program to a different setting [39], and for developing a “clearer understanding of the determinants of successful scaling-up” [11].

Several interviewees suggested that scale-up is more likely through synchronous implementation and research, which Peters and colleagues call “learning and doing” [40]. “Using data and experimenting underlies a lot of successful scale-up approaches,” said one interviewee, who leads health service scale-up projects in LMICs. “Mapping constraints, having the flexibility to redesign, learn-do cycling, being able to call in a more complete set of stakeholders—these kinds of approaches are more likely to lead to success.”